Anda di halaman 1dari 7

Hong Kong Journal of Emergency Medicine

Validation of a Modified Early Warning Score (MEWS) in emergency department observation ward patients

(MEWS) in emergency department observation ward patients TS Lam , PSK Mak , WS Siu ,

TS Lam

in emergency department observation ward patients TS Lam , PSK Mak , WS Siu , MY

, PSK Mak

department observation ward patients TS Lam , PSK Mak , WS Siu , MY Lam ,

, WS Siu

observation ward patients TS Lam , PSK Mak , WS Siu , MY Lam , TF

, MY Lam

ward patients TS Lam , PSK Mak , WS Siu , MY Lam , TF Cheung

, TF Cheung

patients TS Lam , PSK Mak , WS Siu , MY Lam , TF Cheung ,

, TH Rainer

Lam , PSK Mak , WS Siu , MY Lam , TF Cheung , TH Rainer

Objective: The Modified Early Warning Score (MEWS) is a simple physiological scoring system, which can easily be applied at the bedside. The ability of MEWS to identify patients at risk of deterioration in a busy ward was investigated. Method: In a prospective cohort study, we applied MEWS to patients admitted to the 16-bed emergency department observation ward (EDOW) of a tertiary teaching hospital. Results: Data on 427 consecutive EDOW admissions were collected from 7 June to 4 July 2004. Main outcome measures were death, intensive care unit (ICU) admission and inpatient hospital admission. Scores of > 4 were associated with increased risk of death (OR 54.4, 95% CI = 4.7633.7), ICU admission (OR 12.7, 95% CI = 1.1147.3) and hospital admission (OR 9.5, 95% CI = 3.327.9). Conclusion: MEWS is suitable for bedside application in an EDOW setting and may help identify patients at risk of deterioration who require increased levels of care as hospital inpatients and in ICU. Where experienced staff is not available to closely monitor patients in an EDOW, the use of the MEWS system may aid close monitoring and identification of high-risk patients.

(Hong Kong j.emerg.med. 2006;13:24-30)

16 2004 6 7 7 4 427 4 54.4 95% 4.7−633.7 12.7 95% 1.1−147.3 9.5
16
2004
6
7
7
4
427
4
54.4
95%
4.7−633.7
12.7
95%
1.1−147.3
9.5
95%
3.3−27.9

Keywords: Hospital emergency service, observation, physiologic monitoring

emergency service, observation, physiologic monitoring Correspondence to: Timothy Hudson Rainer, MD,

Correspondence to:

Timothy Hudson Rainer, MD, FHKAM(Emergency Medicine)

The Chinese University of Hong Kong, Accident and Emergency Medicine Academic Unit, Shatin, N.T., Hong Kong Email: thrainer@cuhk.edu.hk

Mak Siu Kuen, Paulina, BSc, RN

Prince of Wales Hospital, Accident & Emergency Department, 30-32 Ngan Shing Street, Shatin, N.T., Hong Kong

Lam Tse Sum, MBChB, MRCSEd Siu Wing Sze, MBChB, FHKAM(Emergency Medicine) Lam Mei Yee, MBChB, FHKAM(Emergency Medicine) Cheung Tsz Fung, MBChB, FHKAM(Emergency Medicine)

Introduction

The emergency department (ED) observation ward is the interface between ED care and hospital inpatient or intensive care unit (ICU) management in hospitals in Hong Kong. Because of resource limitations, the number of patients that can be monitored and treated in the ICU and as hospital inpatients is restricted. The identification of patients who might benefit from critical care is therefore crucial.

Lam et al./MEWS and emergency ward patients

The Early Warning Score (EWS) 1 is a tool for bedside evaluation based on five physiological parameters:

systolic blood pressure, pulse rate, respiratory rate, temperature and AVPU score (A for 'alert', V for 'responsive to verbal stimulation', P for 'responsive to painful stimulation', U for 'unresponsive'). 2 The ability of a modified EWS, including relative deviation from patients' normal blood pressure and urine output, to identify surgical patients who would potentially benefit from intensive care, was demonstrated in 2000. 3 The Modified Early Warning Score (MEWS) was validated in medical admissions in 2001. 4 An increasing MEWS score was shown to be associated with worse outcome across a range of specialties, including medicine, surgery and orthopaedics. 5 However, MEWS has not been validated in a heterogeneous group of patients admitted to an ED observation ward (EDOW).

The aims of this study were: (i) to evaluate the ability of MEWS (Appendix) to identify patients at risk; and (ii) to examine the feasibility of MEWS as a screening tool to trigger early assessment and inpatient admission to the hospital or ICU.

Methods

This single-centre, prospective cohort study was conducted in the 16-bed EDOW of a university teaching hospital in the New Territories of Hong Kong from 7 June to 4 July 2004.

After appropriate training, nursing staff collected data while performing routine duties. The following physiological parameters were recorded on admission:

systolic blood pressure, pulse rate, temperature, respiratory rate and either AVPU score or Glasgow Coma Scale (GCS). 6 Blood pressure and pulse rate were measured electronically (Press-mate BP 8800, Colin Electronics, Japan) and checked manually where appropriate. The tympanic temperature was recorded (ThermoScan, Type 6014, Braun, Germany). The respiratory rate was counted over a full minute. AVPU or GCS scores were recorded according to the best response at the time of blood pressure measurement. Nursing staff collected the physiological parameters four times daily (7 am, 11 am, 3 pm and 7 pm) on a

25

dedicated data collection sheet. Completeness of data was checked daily at the bedside by the authors.

The collected data were used to calculate the Modified Early Warning Score (MEWS). A 'critical score' was defined as a MEWS > 4. The highest score reached during EDOW admission was defined as 'ScoreMax'. Primary endpoints were hospital admission to an inpatient ward, ICU admission and death at 30 days. Hospital admission and ICU admission were at the discretion of specialist emergency physicians in the EDOW, who were unaware of the MEWS of the patients.

Statistical analysis was performed using StatView and MedCalc software. The patients eventually admitted to the hospital were compared to those not admitted with regard to their initial physiological parameters by t-tests. We regarded p < 0.05 as statistically significant. Receiver operating characteristic (ROC) curves were generated to identify the performance of each MEWS criterion.

Results

In total, 427 patients were admitted consecutively to the EDOW during the study period. Two patients with incomplete epidemiological or discharge data were excluded, leaving 425 (mean age 59 years; SD = 20; range, 1595; 55% female). Table 1 shows that a heterogeneous group of patients was admitted during the observation period but over 60% had primarily cardiac or gastrointestinal symptoms, or dizziness.

Subsequently 94 patients, already admitted to the EDOW, were assessed by specialist emergency physicians as 'high risk', and were therefore admitted formally to inpatient wards. The specialists had no knowledge of the MEWS at the time. Of these 94 patients, 91 were admitted to the medical ward as 'inpatients' (of whom two died), and three patients were admitted to ICU. MEWS scores in patients admitted to the EDOW ranged from 05, median of 1 (Figure 1). Eleven patients had critical scores (> 4) on initial admission while another six patients developed critical scores after admission to the EDOW (Figure 2).

26

Table 1. Summary of the patients

Hong Kong j. emerg. med. Vol. 13(1) Jan 2006

Variables

n = 425

Age

59 ± 20 years

Sex

Male: 191 (45%);

Female: 234 (55%)

System (emergency department diagnosis or complaints)

Allergic reaction (Angioedema and anaphylaxis)

14 (3%)

Cardiac

98 (23%)

(Chest pain, palpitation, atrial fibrillation, supraventricular tachycardia and congestive heart failure)

Dizziness

78 (18%)

Gastrointestinal (Gastroenteritis, abdominal pain, hepatitis and gastrointestinal bleeding)

87 (20%)

Hyper/hypoglycaemia

11 (3%)

Hypertension

13 (3%)

Neurological (Headache, convulsion and stroke)

19 (5%)

Others (Gout, cellulitis, hypokalemia, heat exhaustion and snake bite)

38 (9%)

Respiratory (Asthma, chronic obstructive pulmonary disease and hyperventilation)

34 (8%)

Trauma (Head injury, sprained back and other minor blunt injuries)

33 (8%)

Observation:

Systolic blood pressure (mmHg)

132 ± 26

Heart rate (bpm)

75 ± 15

Respiratory rate (bpm)

18 ± 2

Temperature (°C)

36.6 ± 0.6

Glasgow Coma Scale score

15 ± 0

Overall ScoreMax

2.3 ± 1.0

Data presented as mean ± SD

Overall ScoreMax 2.3 ± 1.0 Data presented as mean ± SD Figure 1. Distribution of initial

Figure 1. Distribution of initial Modified Early Warning Score (MEWS).

1. Distribution of initial Modified Early Warning Score (MEWS). Figure 2. Distribution of maximum scores (ScoreMax).

Figure 2. Distribution of maximum scores (ScoreMax).

Lam et al./MEWS and emergency ward patients

A ScoreMax > 4 was associated with an increased risk of death (OR 54.4, 95% CI = 4.7633.7), ICU admission (OR 12.7, 95% CI = 1.1147.3) and inpatient hospital admission (OR 9.5, 95% CI 3.3

27.9).

Patients admitted formally to inpatient wards were older and had a higher respiratory rate (Table 2). There were no significant differences with respect to systolic blood pressure, heart rate and temperature.

ROC curves of different physiological parameters and ScoreMax were compared for predicting serious outcome defined as 'death and/or ICU admission' (Figure 3). The area under the curve was highest for ScoreMax with a value of 0.96 (Table 3). ROC curves were also compared for hospital admission (Figure 4). The area under the curve was highest for respiratory rate with a value of 0.77 (Table 3).

27

The sensitivity and specificity of ScoreMax for death or ICU admission were calculated using a ROC curve of ScoreMax and serious outcome (death/ICU). Criteria > 3 performed best, yielding a sensitivity of 100% (95% CI = 48100%) and a specificity of 88% (95% CI = 8591%). Using criteria of > 4 yielded a sensitivity of 60% (95% CI = 1594%) and specificity of 97% (95% CI = 9598%).

Discussion

The Modified Early Warning Score is best regarded as a defined judgement on routinely recorded physiological data. Using previously published scoring criteria, 1,3,4 this study has demonstrated that a raised MEWS was associated with increased mortality in a group of EDOW admissions. Calculation of the MEWS for EDOW admissions might be useful to

Table 2. Physiological parameters on admission of patients requiring or not requiring admission to hospital

 

No admission

Admission

p-value

Number Age (years) Systolic BP (mmHg) Pulse rate (bpm) Respiratory rate (bpm) Temperature (Celsius)

331

94

57 ± 20

64 ± 18

0.005

139 ± 28

140 ± 28

0.76

81 ±

17

83 ±

18

0.32

18 ±

2

20 ±

3

< 0.0001

36.6 ± 0.6

36.8 ± 0.9

0.12

Data presented as mean ± SD, p-value for independent samples t-test

Table 3. Area under ROC curve of each parameter

Parameter

Serious outcome

Admission

ScoreMax

0.96

0.73

Respiratory rate

0.91

0.77

Heart rate

0.89

0.55

Temperature

0.79

0.51

Systolic blood pressure

0.61

0.51

Glasgow Coma Scale

0.50

0.50

0.61 0.51 Glasgow Coma Scale 0.50 0.50 Figure 3. ROC curves for serious outcome (death/ICU

Figure 3. ROC curves for serious outcome (death/ICU admission).

28

28 Figure 4. ROC curves for hospital admission. identify patients at highest risk of deterioration. Appropriate

Figure 4. ROC curves for hospital admission.

identify patients at highest risk of deterioration. Appropriate interventions could then be targeted upon a small number of at-risk patients among the 10-20 daily EDOW admissions.

Subbe suggested that case mix might be one important factor. 4 Application of MEWS in the EDOW demonstrated that it was useful in a heterogeneous group of patients.

To prioritise scarce hospital resources, it would be valuable to identify patients who would benefit from inpatient hospital or ICU admission, as well as those whose inpatient or ICU admission could be prevented by changes of management in the EDOW. As patients with critical scores (> 4) in this study were at increased risk of catastrophic deterioration, MEWS might be a helpful screening tool to triage patients for intensified treatment in a hospital ward or ICU rather than in an EDOW.

By comparing MEWS to predict inpatient hospital admission and death or ICU admission, it was found that MEWS was more sensitive and specific in predicting death or ICU admission over hospital ward admission. MEWS is thus more useful with serious patients.

Previous work showed no benefit of introducing a MEWS system on clinical outcomes, cardiopulmonary

Hong Kong j. emerg. med. Vol. 13(1) Jan 2006

arrests and ICU utilisation in acute medical admissions. 7 It may be useful to devise a protocol in the EDOW with standardised response to abnormal values, and investigate the effect afterwards. Although we had used a critical score > 4 in this study to calculate odds ratios, we demonstrated that scores > 3 were more sensitive in predicting serious outcome. This exploratory study sought to identify the trigger score for adverse outcomes, which was found to be MEWS > 3. This criterion should be the trigger score for immediate specialist emergency physician review when MEWS is clinically implemented. Future studies may help to identify and confirm the most appropriate cut-off score for different at-risk populations.

Respiratory rate was demonstrated to be the best discriminating physiological parameter to predict serious outcome (death or ICU admission). Respiratory rate is less commonly documented compared to blood pressure, heart rate, GCS and temperature during routine observation. Although taking the respiratory rate is more labour-intensive than the automated techniques of measuring blood pressure, heart rate and temperature, it is a recommended parameter for routine observations. The long-term beneficial effect of introducing the MEWS system on respiratory rate recording in general wards has recently been demonstrated. 8

Unlike previous studies, GCS was not a predictor in our study. This was most likely due to the fact that patients with decreased GCS would not be admitted to our EDOW. Unlike some Western societies where excessive alcohol ingestion is a major problem, such alcohol consumption is rare in our setting, and the few patients who present with alcohol-related decreased GCS are observed close to the resuscitation room rather than the EDOW.

Figure 3 shows that although respiratory rate was the best individual predictor among the parameters measured, the MEWS performed best overall. In addition, although the difference in mean respiratory rate was statistically significant, clinically the difference

Lam et al./MEWS and emergency ward patients

was not as great (20 vs. 18 bpm) and therefore respiratory rate might not be suitable as a sole indicator of serious illness.

Our study was limited by several factors. It was a single- centre study on a limited number of patients in a specific local setting. The majority of patients who were admitted as inpatients to the hospital or ICU, or those who died, would have had improvements and deteriorations following transfer out of the EDOW. MEWS-type data leading up to those events would probably give additional information regarding physiology prior to catastrophic events. Moreover, around one-third of the EDOW daily admissions occurred between 7 pm and 7 am (n = 5) and this group was at higher risk of unnoticed deterioration if MEWS was not calculated during that period. If MEWS were to be implemented generally in clinical practice, further observation would need to be done at 11 pm and 3 am to ensure patient monitoring was optimal.

The implementation of MEWS was acceptable to the nursing staff, although thought to be more labour intensive and complex than recording simple observations. Some nursing staff members were not convinced that MEWS was an improvement on their clinical judgement and experience. As patients in the study were not randomised to two groups those

29

evaluated by MEWS and those evaluated by clinical judgement alone it is not possible to say to what extent implementing MEWS will improve patient surveillance and decision-making.

Future studies may include oxygen saturation (SpO 2 ) as one of the physiological parameters to see if it behaves similar to respiratory rate as a potential discriminator to identify patients at risk. It may also be useful to evaluate MEWS for patients in the emergency department 'majors' area which has a distinct population from that observed in the EDOW.

Conclusion

The MEWS is able to identify patients at risk and is feasible as a screening tool in the setting of unselected EDOW admissions to trigger early assessment and inpatient admission to the hospital or ICU.

Acknowledgements

We thank all nurses of the emergency department, Mr HS Lam and Ms Katherine Lam for help with data collection; Dr. KL Ong, Dr. TW Wong and Professor CA Graham for their helpful comments; and all others who helped to make this possible.

Appendix. Modified Early Warning Score

MEWS

+3

+2

+1

0

+1

+2

+3

Systolic blood pressure Heart rate Respiratory rate Temperature AVPU/GCS score

< 70

7080

81100

101199

200

< 40

4150

51100

101110

111130

> 130

< 9

914

1520

2129

30

< 35

35.136

36.138

38.138.5

>38.5

< 9

913

14

A/15

V/Confused

P

U

The score is calculated by measuring the five parameters as above and adding together the assigned score for each physiological value. AVPU = Alert, Verbal, Pain, Unresponsive; GCS = Glasgow Coma Scale

30

References

1. Morgan RJM, Williams F, Wright MM. An Early Warning Scoring System for detecting developing critical illness. Clin Intens Care 1997;8:100.

2. American College of Surgeons Committee on Trauma. Advanced Trauma Life Support Course for Physicians. Chicago, IL: American College of Surgeons; 1993.

3. Stenhouse C, Coates S, Tivey M, Allsop P, Parker T. Prospective evaluation of a modified Early Warning Score to aid detection of patients developing critical illness on a surgical ward. Br J Anaesth 2000;84:663.

4. Subbe CP, Kruger M, Rutherford P, Gemmel L. Validation of a modified Early Warning Score in medical admissions. QJM 2001;94:5216.

Hong Kong j. emerg. med. Vol. 13(1) Jan 2006

5. Quarterman CP, Thomas AN, McKenna M, McNamee R. Use of a patient information system to audit the introduction of modified early warning scoring. J Eval Clin Pract 2005;11(2):1338.

6. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet 1974;2:814.

7. Subbe CP, Davies RG, Williams E, Rutherford P, Gemmel L. Effect of introducing the Modified Early Warning Score on clinical outcomes, cardio-pulmonary arrests and intensive care utilisation in acute medical admissions. Anaesthesia 2003,58(8):797802.

8. McBride J, Knight D, Piper J, Smith GB. Long-term effect of introducing an early warning score on respiratory rate charting on general wards. Resuscitation

2005;65:414.